Does Prompt Compliance with the COSO 2013 Framework Signal a Commitment to a Strong Internal Control Environment?
Abstract
In this study, we investigate the determinants of compliance with the Committee of Sponsoring Organizations of the Treadway Commission (COSO) 2013 framework and whether prompt compliance provides a signal of a commitment to a strong internal control environment. COSO 2013 framework represents the biggest change to the internal control framework in more than two decades. In firms’ first fiscal year following the supersession of the COSO 1992 framework, only 91 percent of firms in our sample were in compliance with the updated COSO 2013 framework. We find that compliance with the updated framework is more likely among firms that are larger, older, more highly leveraged, less complex, that operate in more litigious industries, and that have an effective internal control environment. Controlling for potential selection bias, we next examine whether compliance with the updated framework is indicative of a higher level of control consciousness and governance as evidenced by more conservative financial reporting. Finally, we use short-window market reactions to quarterly earnings surprises to examine whether investors perceive compliance with the updated framework as an indication of the overall control consciousness and governance of the firm. We find that firms that comply with the COSO 2013 framework exhibit more conservative financial reporting and that investors react more positively to these firms’ quarterly earnings surprises following initial compliance. Importantly, these results hold among a sample of firms without reported material weaknesses in internal controls. These results provide evidence that firms can help alleviate agency costs by signaling their commitment to a strong internal control environment.
Keywords: COSO2013 internal control framework; Accounting conservatism; prompt
1
I. INTRODUCTION
A long stream of research, beginning with Jensen and Meckling (1976), suggests that agency problems can arise in corporate settings when there is greater information asymmetry between managers and shareholders. Prior research also suggests that corporate internal control systems can play an important governance role in monitoring managerial behavior and reducing agency costs (e.g., Doyle et al. 2007a; Goh and Li 2011). This is not only evidenced by lower earnings quality among firms with weak internal control systems (Doyle et al. 2007a), but also by evidence of negative investor reactions to disclosures of such weaknesses (Hammersley et al. 2008). Although the absence of disclosed internal control weaknesses provides some indication of the effectiveness of a firm’s control environment, disclosure is limited to known weaknesses that could materially impact the reported numbers as of the financial reporting date. No
disclosure is required for material weaknesses existing during the reporting period but remediated before the period end date. Additionally, prior research suggests that a large proportion of firms with material
weaknesses in internal controls fail to report in a timely manner (Rice and Webber 2012). To help alleviate agency costs, managers may make intentional decisions to signal their commitment to a strong control environment.
In this study, we examine whether prompt compliance with the most up to date internal control framework signals a commitment to a strong internal control environment. Such a commitment is likely to manifest itself in more effective internal controls and more conservative financial reporting (Garcia et al. 2009; Goh and Li 2011). Specifically, we investigate determinants of prompt compliance with the Committee of Sponsoring Organizations of the Treadway Commission (COSO) 2013 framework using several factors likely associated with the strength of a firm’s overall control environment. We then perform several analyses to investigate whether firms that promptly comply with COSO 2013 provide more conservative financial reporting and whether
2
investors perceive compliance with the updated framework as an indication of a firm’s commitment to a strong system of internal control.
COSO released its original internal control framework in 1992, which became the basis for auditors to assess and report on their clients’ internal control over financial reporting under the provisions of the Sarbanes-Oxley Act of 2002 (SOX). However, since the release of this original framework, businesses and operating environments have changed dramatically. In response to these changes, COSO introduced its updated internal control framework, also known as the 2013 internal control framework, on May 14, 2013 (hereafter referred to as the COSO 2013 framework). During the transition period from May 14, 2013 to December 15, 2014, public firms and their auditors had the choice to use either the original 1992 framework or the updated 2013 framework as the underlying basis for their internal control assessment as long as the applicable framework was disclosed. Following the transition period, however, COSO considers the 1992 framework as having been superseded. Despite this, in firms’ first fiscal year following supersession of the COSO 1992 framework, we find that only 91 percent of firms in our sample comply with the updated COSO 2013 framework.
We perform our tests using a sample of firms subject to the reporting requirements of Section 404(b) of SOX, which requires auditor attestation on the effectiveness of internal control over financial reporting, with fiscal years ending after December 15, 2014 through May 31, 2016. Because our sample period begins after the transition period, firms using the COSO 2013 framework are identified as “compliance” firms. We refer to “noncompliance” firms as those that continue to use the COSO 1992 framework. Our sample consists of 3,564 firms that use the updated COSO 2013 framework, and 347 firms that continue to use the original 1992 framework. Building on prior research, (Ge and McVay 2005; Doyle et al. 2007b; Feng et al. 2015), we examine the determinants of compliance with the COSO 2013 framework based on variables associated with the strength of a firm’s control environment, which
3
include the disclosure of material weaknesses in internal controls, firm size, leverage, age, complexity, the occurrence of significant or unusual transactions, financial distress, whether the firm operates in a litigious industry, and the risk of financial reporting improprieties. We find that compliance is more likely among firms that do not disclose internal control weaknesses, are larger, more highly leveraged, older, less complex, are involved in more merger and acquisition activity, have restructuring charges, and operate in a more litigious industry.
Next, we examine the relation between compliance with the COSO 2013 framework and financial reporting conservatism. We employ three measures of conservatism: C_SCORE developed by Khan and Watts (2009), CON_ACC suggested by Givoly and Hayn (2000) and Ahmed et al. (2002), and the asymmetric timeliness of earnings following Basu (1997). Because compliance with the COSO 2013 framework is not randomly determined, we address the possibility of selection bias by estimating a two-stage Heckman selection model. We use the determinants model discussed above as our first stage model. To satisfy the exclusion restriction, we use the indicator variable for whether the firm is a client of KPMG.1 From this model, we derive the inverse Mills ratio to control for the unobservable factors associated with the decision to comply with the COSO 2013
framework. After controlling for known determinants of accounting conservatism and for potential self-selection bias, we find a positive association between firms complying with the COSO 2013 framework and all three measures of conservatism. These findings are consistent with our hypothesis that firms promptly complying with the COSO 2013 framework exhibit a commitment to control
consciousness and strong governance.
1 Based on discussions with a COSO board member, KPMG did not encourage early compliance with the updated framework. We believe this variable meets the criteria of an exclusion restriction variable as it should affect the decision to comply, but should not necessarily affect the outcome variables of our second stage models. We include an indicator variable for whether the firm’s auditor is KPMG and find that this variable negatively predicts the likelihood of compliance with the COSO 2013 framework.
4
We next examine whether investors perceive compliance with the COSO 2013 framework as an indication of a higher level of control consciousness and governance, as evidenced by more positive short-window market reactions to quarterly earnings surprises. We find that firms
complying with the COSO 2013 framework experience more positive short-window reactions to quarterly earnings surprises, suggesting that investors perceive these firms’ earnings to be more informative and of higher quality. In further analysis, we incorporate pre-implementation quarterly observations beginning with firms’ second fiscal quarter of 2013 and use a difference-in-difference estimation to determine whether the investor reaction to quarterly earnings surprises is incrementally higher in the quarters following disclosure of COSO 2013 compliance. We document a higher earnings response coefficient in the quarters following disclosure of COSO 2013 compliance, providing further evidence that prompt compliance with the updated framework signals the firm’s overall commitment to control consciousness and governance.
In additional analyses, we examine whether these relations hold when limiting our sample to firms without disclosed internal control weaknesses. If prompt compliance with the updated COSO framework does indeed provide an indication of a strong commitment to internal control and governance, then we would expect to observe a consistent result among firms with no reported weaknesses in internal controls for which investors have less ability to differentiate variation in the strength of the internal control environment. Consistent with our expectation, we find that among firms with no reported weaknesses in internal control, prompt compliance with the COSO 2013 framework is associated with greater financial reporting
conservatism and more positive market reactions to quarterly earnings surprises.
This study contributes to the literature in several ways. First, the results of this study highlight that a non-trivial amount of companies did not comply with the COSO 2013 framework following supersession of the COSO 1992 framework. To our knowledge, this is the first study that examines the determinants of
5
compliance with the COSO 2013 framework. Second, this study contributes to the literature by examining whether prompt compliance with an updated internal control framework can serve as a signal of a strong system of internal controls (e.g., Doyle et al. 2007a; Goh and Li 2011) to help alleviate agency costs. The results support this notion. Specifically, compliance with the COSO 2013 framework is associated with more conservative financial reporting and improves investors’ perceptions of the underlying quality of the earnings numbers. These results highlight the importance of prompt compliance with future framework updates, new rules, or new regulation involving internal control over financial reporting.
The remainder of this paper is organized as follows. Section II reviews prior literature and develops the hypotheses. Section III describes sample selection and research design. Section IV presents our main results. Section V provides additional analyses. Section VI concludes the paper.
II. BACKGROUND, PRIOR LITERATURE, AND DEVELOPMENT OF HYPOTHESES
COSO’s 2013 Internal Control Framework
On May 14, 2013, COSO released its updated internal control framework, known as the COSO 2013 framework.2 At the time, the COSO Board considered it proper for public companies to continue to use their original 1992 framework during the transition period between May 14, 2013 and December 15, 2014 (COSO 2013). During this period, companies and their auditors were required to clearly disclose which framework they used. Following the transition period, the original 1992 framework is considered superseded by COSO. The COSO 2013 framework is similar to the original 1992 framework, but provides several significant changes. Although the five components of a firm’s internal control system – control environment, risk assessment, control activities, information and communication, and monitoring activities – remain intact, the updated framework provides “explicit articulation of the 17 principles” that are meant to codify the fundamental concepts related to those five components (COSO 2013). In addition to these
6
relevant 17 principles, the COSO 2013 framework introduces 81 points of focus which provide greater detail and insight into the principles. The primary goal of the updated COSO 2013 framework is to enhance assessment and evaluation to determine if the explicitly stated principles are present and functioning (i.e., less ambiguity), reflect increased relevance and use of technology and related controls, incorporate enhanced discussion of governance concepts, enhance anti-fraud expectations, and increase the focus on non-financial reporting objectives.
Determinants of Compliance with COSO 2013 Framework
To date, there is no research that examines the determinants of compliance with the COSO 2013 framework. If prompt compliance is indicative of firms’ commitment to strong governance and a strong internal control environment then we would expect that the likelihood of compliance will vary based on factors associated with the strength of a firm’s internal control. Doyle et al. (2007b) investigate several potential firm characteristics that determine internal control material weaknesses using 779 firms that disclose material weaknesses from August 2002 to 2005. Relying on prior research findings, we examine whether these firm characteristics, which include the disclosure of material weakness in internal control, firm size, leverage, age, complexity, the occurrence of unusual or significant transactions, financial distress, litigation risk, and the risk of misstatement, are associated with compliance with the COSO 2013
framework. We elaborate on each of these firm characteristics below. Because the disclosure of a material weakness in internal controls is a clear indication of a weak internal control
environment, we argue that prompt compliance with the COSO 2013 framework will be less likely for these firms. Our first hypothesis (stated in alternative form) is as follows:
H1a: Ceteris paribus, firms with disclosed material weaknesses in internal controls are less
likely to comply with the COSO 2013 framework immediately following the transition period.
7
Apart from the disclosure of a material weakness in internal controls, other firm
characteristics should be associated with the overall internal control environment of the firm. For example, firm size has been shown to be associated with stronger internal control systems (Ge and McVay 2005; Doyle et al. 2007b). Smaller firms tend to have less sophisticated internal control systems and fewer resources than larger firms (e.g., fewer investments in technology or less experienced or competent staff). Smaller firms are also less likely to benefit from economies of scale when they manage and operate their internal control systems (Doyle et al. 2007b). Furthermore, smaller firms are more likely to have limited time and resources to monitor their internal control system. The adoption and implementation of the COSO 2013 framework requires dedicated resources to identify relevant changes necessary to internal controls, update existing documentation, and ensure compliance with the new framework. Smaller firms with greater resource constraints are likely at a disadvantage. This leads to our next hypothesis (stated in alternative form):
H1b: Ceteris paribus, larger firms are more likely to comply with the COSO 2013 framework
immediately following the transition period.
We also consider a firm’s financial leverage as a potential factor affecting compliance with the COSO 2013 framework. On the one hand, highly leveraged firms may not have
sufficient resources and funds to allocate to the adoption and implementation of the COSO 2013 framework. On the other hand, firms with higher leverage may have sufficient (or more) cash to acquire the resources necessary to promptly comply with the new framework. As such, we do not make a directional prediction on the association between leverage and compliance with the COSO 2013 framework. Our next hypothesis (stated in null form) is as follows:
H1c: Ceteris paribus, compliance with the COSO 2013 framework immediately following the
8
Firm age is also a potential determinant of compliance with the COSO 2013 framework. Because younger firms are growing and changing more rapidly, these firms may not have resources to establish or invest in a sophisticated internal control system (Feng et al. 2015). Additionally, they may be reluctant to invest the time and resources necessary for prompt compliance. Thus, we expect that older firms are more likely to comply with the COSO 2013 framework, which leads to the following hypothesis (stated in the alternative):
H1d: Ceteris paribus, older firms are more likely to comply with the COSO 2013 framework
immediately following the transition period.
Firm complexity could influence prompt compliance with the COSO 2013 framework. Doyle et al. (2007b) find that firm complexity increases the likelihood of disclosing a material weakness in internal controls. Complex firms may need more time to assess and document compliance with the updated internal control framework. Thus, we expect that a more complex firm will be less likely to promptly comply with COSO 2013 framework. This leads to the following hypothesis (stated in the alternative):
H1e: Ceteris paribus, more complex firms are less likely to comply with the COSO 2013
framework immediately following the transition period.
We also consider whether the occurrence of significant or unusual transactions during the year the updated framework becomes effective. Restructuring and merger and acquisition
activities are a time-consuming process for a firm. Not only will these types of events affect the structure of the firm and its related internal controls and processes, but they will divert the time and attention of the accounting and financial reporting staff that would also likely be responsible for assessing and updating documentation related to compliance with the COSO 2013
framework. These types of activities have been shown to increase the likelihood of material weaknesses in internal control (Doyle et al. 2007b). Thus, we expect that firms undergoing
9
restructuring and involved in merger and acquisition activity are more likely to delay compliance with the COSO 2013 framework. This leads to the following hypothesis (stated in the
alternative):
H1f: Ceteris paribus, compliance with the COSO 2013 framework immediately following the
transition period is less likely among firms experiencing restructuring or merger and acquisition activity.
Additionally, firms in financial distress may lack the necessary resources to dedicate to prompt compliance with the updated framework. As such, we expect that firms experiencing financial distress are less likely to comply with the COSO 2013 framework. This leads to our next hypothesis (stated in alternative form):
H1g: Ceteris paribus, financially distressed firms are less likely to comply with the COSO 2013
framework immediately following the transition period.
Finally, we examine whether litigation risk or misstatement risk affect the likelihood of complying with the updated framework. Firms in litigious industries may be more inclined to adhere to industry norms to avoid potential litigation. Although compliance with the COSO 2013 framework is not enforced by a regulatory authority, managers may believe that noncompliance would increase the risk of litigation. Firms with greater misstatement risk may have weaker internal controls. If the control environment is weak, prompt compliance may be less likely. As such, we hypothesize the following (stated in the alternative):
H1h: Ceteris paribus, firms in litigious industries are more likely to comply with the COSO
2013 framework immediately following the transition period; and
H1i: Ceteris paribus, firms with greater misstatement risk are less likely to comply with the COSO
2013 framework immediately following the transition period.
COSO 2013 Framework Compliance and Accounting Conservatism
10
(COSO 2013) and should enable organizations to effectively design, implement, and reevaluate their internal controls associated with operating, reporting, and compliance objectives (COSO 2013). Prior research suggests that effective internal controls can act as an important corporate governance mechanism in monitoring managerial behavior and thus mitigating agency problems (Jensen 1993). Consistent with this, prior studies document that firms with internal control material weaknesses have lower financial reporting quality (Doyle et al. 2007a; Ashbaugh-Skaife et al. 2008), higher information risk (Beneish et al. 2008; Ashbaugh-Skaife et al. 2009), and less accurate management guidance (Feng et al. 2009). Furthermore, Goh and Li (2011) find that internal control quality is positively related to conditional accounting conservatism. They find that firms with material weaknesses in internal controls exhibit less conservative financial reporting. They also find that firms that subsequently remediate material weaknesses exhibit greater conservatism compared to firms that continue to report material weaknesses in internal controls. Prior studies also highlight how conservative policies and choice on accounting matters helps alleviates agency conflicts between mangers and shareholders (Holthausen and Watts 2001; Ahmed et al. 2002; Watts 2003a; Ball and Shivakumar 2006). Consistent with this, Garcia et al. (2009) find a positive association between conditional accounting conservatism and strong corporate governance. The results of these studies suggest that conservative financial reporting is indicative of a commitment to control consciousness and governance.
To help alleviate agency costs, managers may promptly comply with the updated framework to signal the firm’s commitment to strong governance and a strong internal control environment. As such, we expect that prompt compliance with the COSO 2013 framework is positively associated with accounting conservatism. This leads to the following hypothesis (stated in alternative form):
H2: Ceteris paribus, firms that comply with the COSO 2013 internal control framework after
11
COSO 2013 Framework Compliance and Investor Perceptions of Earnings Quality
If compliance with the COSO 2013 framework is an intentional signal of management’s commitment to a strong internal control environment then investors should find these firms’ earnings more informative relative to noncompliance firms. Prior research finds that investors react negatively to internal control weakness disclosures (Hammersley et al. 2008) suggesting increased risk or uncertainty in that firm’s future earnings. Additionally, firms’ earnings response coefficients increase after the improvements or enhancements to governance and internal controls. For example Chan et al. (2012) find increased earnings response coefficients following adoption of compensation clawback provisions. Given that compliance with the COSO 2013 framework is a potential signal of a firm’s commitment to a strong internal control environment and higher financial reporting quality, we hypothesize that market participants find quarterly earnings surprises more informative for these firms relative to the earnings surprises of noncompliance firms, particularly after the initial disclosure of compliance. This leads to the following hypothesis (stated in alternative form):
H3: Ceteris paribus, firms that comply with the COSO 2013 internal control framework after
the transition period have higher quarterly earnings response coefficients than noncompliance firms, which are incrementally more informative following initial disclosure of compliance.
III. SAMPLE AND RESEARCH DESIGN
Sample Selection
To obtain the sample, we use the Audit Analytics database to identify whether firms disclose compliance with the COSO 2013 framework in the first fiscal year following the effective date of the updated framework (i.e., firms with fiscal years ending after December 15, 2014 through May 31, 2016). We obtain data related to internal control weaknesses and going concern opinions, from Audit Analytics.
12
We obtain annual financial data from the Compustat annual database and monthly stock return data from the CRSP database. We collect segment data from the Compustat Segment file. We merge these datasets and remove firms in financial industries (SIC codes 6000-6999) (Ahmed and Duellman 2013). We allow our sample sizes to vary slightly based on data availability for the variables in the respective models.
Empirical Models
Determinants of Compliance with COSO’s 2013 Framework
We first examine the determinants of compliance with the COSO 2013 framework using a sample of firms with fiscal years ending after December 15, 2014 through May 31, 2016. As discussed above, we model the likelihood of prompt compliance with the COSO 2013 framework as a function of firm characteristics associated with the strength of a firm’s internal control. To do this, we estimate the following logistic regression model:
Prob (Compliancei,t =1) = F(β0 +β1ICMWi,t + β2Sizei,t + β3Leveragei,t + β4Firm_Agei,t
+ β5Segmentsi,t + β6Foreign_Operationsi,t + β7M&Ai,t
+ β8Restructuringi,t + β9Aggregate_Lossesi,t + β10Going_Concerni,t + β11Litigationi,t + β12F_SCOREi,t + β13KPMGi,t), (1)
where Compliance is an indicator variable equal to one if a firm uses the COSO 2013 internal framework after December 15, 2014 and zero otherwise. We build on prior research findings examining determinants of internal control weaknesses (Doyle et al. 2007a), and examine whether compliance with the updated framework is a function of disclosure of material
weaknesses in internal control, firm size, leverage, age, complexity, the occurrence of unusual or significant transactions, financial distress, litigation risk, and the risk of misstatement. We measure firm size (Size) using the natural logarithm of total assets. We measure leverage
(Leverage) as the sum of short-term and long-term debt divided by total assets. We measure firm age (Firm_Age) using the natural logarithm of the number of years a firm has data on the CRSP
13
database. We capture firm complexity using the natural logarithm of the sum of the number of operating and geographic segments (Segments) and an indicator variable for whether the firm has foreign operations (Foreign_Operations). We capture significant or unusual transactions using an indicator variable for whether a firm has restructuring charges (Restructuring) and an indicator variable for whether a firm is engaged in a merger or acquisition during the year (M&A). We capture financial distress using an indicator variable if the firm reports a loss in the current and prior year (Aggregate_Losses) and an indicator variable if the firm received a going-concern audit report modification (Going_Concern). Following Francis et al. (1994), we use an indicator variable for firms in litigious industries (i.e., SIC codes 2833–2836, 3570–3577, 3600– 3674, 5200–5961, and 7370) (Litigation). Following Dechow et al. (2011), we use the F-score to capture misstatement risk (F_SCORE).3 Finally, we include an indicator variable if the firm’s auditor is KPMG (KPMG) based on discussions with a COSO board member suggesting that KPMG did not encourage early compliance with the updated framework. Additionally, we control for industry fixed effects in equation (1) and use robust standard errors clustered at the firm level. All variables are defined in Appendix A.
Compliance with COSO’s 2013 Framework and Accounting Conservatism
We next investigate the relation between compliance with the COSO 2013 framework and accounting conservatism. To do this, we estimate the following OLS regression model: ACC_Conservatismi,t = β0 + β1Compliancei,t + β2Sizei,t + β3Leveragei,t
+ β4Market-to-Booki,t + β5ROAi,t + β6Firm_Agei,t + β7Sales_Growthi,t
+ β8Rd_Advi,t + β9Litigationi,t + β10Big4i,t + β11Inverse_Mills_Ratioi,t
+ Σi,tIndustry_Dummy+ ɛi,t, (2)
14
where ACC_Conservatismrepresents three different dependent variables, C_SCORE, CON_ACC, and CON_SKEWNESS. We estimate C_SCORE, a measure of conditional
conservatism, following Khan and Watts (2009) (See Appendix B for more details). Larger values of C_SCORE indicate greater conditional accounting conservatism. Additionally, we use a measure of unconditional conservatism, CON_ACC, following Givoly and Hayn (2000), Ahmed et al. (2002), and Ahmed and Duellman (2013). We calculate CON_ACC as net income before extraordinary items plus depreciation expense minus cash flow from operations, deflated by average total assets, averaged over a 3-year period centered on year t. We multiply this measure by negative one so that higher values of CON_ACC mean greater unconditional accounting conservatism. Finally, we use another unconditional conservatism measure,
CON_SKEWNESS, following Givoly and Hayn (2000) and Ahmed and Duellman (2013). This
measure is calculated as the difference between cash flow skewness and earnings skewness. The skewness of cash flow (earnings) is defined as (𝑥 − 𝜇)3/𝜎3 where 𝑥 is cash flows (earnings), and 𝜇 and 𝜎 are the mean and standard deviation of cash flows (earnings) over the last five years. Higher values of CON_SKEWNESS indicate greater unconditional accounting conservatism. We control for industry fixed effects in equation (2) and use robust standard errors clustered at the firm level.
The variable of interest in equation (2) is an indicator variable, Compliance, which we predict to be positive. Based on prior studies (Ahmed et al. 2002; Ahmed and Duellman 2007, 2013; Givoly et al. 2007; Roychowdhury and Watts 2007; LaFond ad Roychowdhury 2008; LaFond and Watts 2008; Goh and Li 2011; Zhang 2012), we control for firm characteristics and external auditor characteristics that have been shown to affect accounting conservatism.
15
firms have less information asymmetry, thereby decreasing the demand for conservatism. In contrast, as Watts and Zimmerman (1978) argue, large firms are more likely to face large political costs so that they engage in more conservative accountings. Given the two sides of the argument, we expect no relationship between ACC_Conservatism and Size. We also include firm leverage (Leverage) as a control variable. Ahmed et al. (2002) argue that highly leveraged firms tend to have more conservative accounting due to their greater bondholder-shareholder conflicts. Following Roychowdhury and Watts (2007) and Ahmed and Duellman (2013), we control for the market-to-book ratio (Market-to-Book). Ahmed and Duellman (2013) find a negative relation between accounting conservatism and the market-to-book ratio. Firms with high market-to-book ratio are likely to have more growth opportunities, thereby increasing information asymmetry between managers and investors. Thus, there will likely be an increase in demand for accounting conservatism (LaFond and Watts 2008). We control for return on assets (ROA) because Ahmed et al. (2002) document that firms with a higher return on assets choose more conservative accounting. We include firm age (Firm_Age) because Khan and Watts (2009) predict a negative relation between firm age and accounting conservatism. We expect that accounting conservatism decreases with firm age. Following Ahmed et al. (2002), we control for sales growth
(Sales_Growth). Ahmed et al. (2002) and Ahmed and Duellman (2007, 2013) find that sales growth is negatively related to the accrual-based conservatism measure, CON_ACC. We expect a negative relation between CON_ACC and Sales_Growth. We control for research and
development (R&D) and advertising expenditures (Rd_Adv) as Ahmed and Duellman (2007, 2013) argue that firms with high R&D and advertising expenditures use more conservative accounting. Following Basu (1997) and Watts (2003a), we control for whether a firm is a member of a litigious industry (Litigation). We also control for whether a firm is audited by a
16
Big 4 auditor (Big4). We expect firms with a Big 4 auditor to be more conservative. We include the inverse Mills ratio (Inverse_Mills_Ratio) from the determinants model in equation (1) to control for the unobservable factors associated with a firm’s decision to comply with the COSO 2013 framework. Finally, as in Givoly et al. (2007), we include industry fixed effects to control for variation in accounting conservatism across industries. All variables are defined in Appendix A.
Our third proxy for accounting conservatism is based on the Basu (1997) asymmetric timeliness measure. Building on the Basu (1997) specification, we estimate the following regression model:
NIi,t = β0 + β1DRi,t + β2Compliancei,t + β3Sizei,t + β4Leveragei,t +Β5Market-to-Booki,t + β6Litigationi,t +β7DRi,t*Compliancei,t + β8DRi,t* Sizei,t + β9DRi,t*Leveragei,t + β10DRi,t*Market-to-Booki,t + β11DRi,t*Litigationi,t + β12RETi,t +β13RETi,t*Compliancei,t
+ β14RETi,t*Sizei,t +β15RETi,t*Leveragei,t + β16REi,t*Market-to-Booki,t + β17RETi,t *Litigation+ β18DRi,t*RETi,t + β19DRi,t*RETi,t*Compliancei,t +β20DRi,t*RETi,t*Sizei,t + β21DRi,t*RETi,t*Leveragei,t
+ β22DRi,t*RETi,t*Market-to-Booki,t + β23DRi,t*RETi,t*Litigationi,t
+Σi,tIndustry_Dummy+ ɛi,t, (3)
where NIis defined as the earnings before extraordinary items divided by the market value of equity at the beginning of the fiscal year. DRis an indicator variable equal to one if a firm’s return is negative and zero otherwise. RETis the buy-and-hold return over the fiscal year. The variable of interest in equation (3) is β19, the coefficient on the triple interaction of DR, RET, and
Compliance, which we predict to be positive. As in equation (2), we control for industry fixed
17
Compliance with COSO’s 2013 Framework and Investor Perceptions of Earnings Quality
We next investigate whether compliance with the COSO 2013 framework affects investor perceptions about the quality of the firm’s earnings. To do this, we examine whether the short-window earnings response coefficient (ERC) to quarterly earnings surprises is more positive for compliance firms. To do this, we estimate the following OLS regression model following Francis and Ke (2006) and Ghosh et al. (2009):
CAR = β0 + β1FERRq + β2Compliance + β3FERRq*Compliance+ βX + βIndustry FE+
βQuarter-Year FE + βFERRq*X+ βFERRq*Industry FE+ βFERRq*Quarter-Year
FE + ɛit (4)
where CAR is the abnormal (i.e., market-adjusted) returns cumulated over days [-1, +1] relative to the quarterly earnings announcement. FERR is the analyst forecast error, measured as the difference between reported quarterly earnings per share and the most recent median consensus analyst earnings forecast, deflated by prior quarter stock price. X is a vector of control variables following prior research (Francis and Ke 2006; Ghosh et al. 2005; Ghosh et al. 2009), which includes the absolute value of FERR (absFERR), an indicator variable if net income for the quarter is less than zero (Loss), an indicator variable if special items is five percent or more of total assets (Restructure), the ratio of short and long-term debt to total equity (DE), an indicator variable for the last fiscal quarter in the respective year (QTR4), the natural log of the market value of equity (LnMV), and the standard deviation of market-adjusted buy-and-hold returns (STD_Return).4 Finally, we include industry and quarter-year fixed effects to control for variation in short-window cumulative abnormal returns across industries and over time.
4 Market-adjusted are calculated as the difference between raw returns and the value-weighted market returns from the CRSP database over the previous 60 months.
18
Consistent with prior research, the ERC is the coefficient on FERR. The variable of interest in equation (4) is β3, the coefficient on the interaction of FERR and Compliance, which
we predict to be positive. A positive coefficient would indicate that market participants find the earnings of compliance firms more informative than the earnings of noncompliance firms.
To strengthen inferences that prompt compliance with COSO 2013 serves as a signal to market participants, we include in the sample pre-implementation quarterly observations beginning with firms’ second fiscal quarter of 2013 through May 2016 and use a difference-in-difference estimation with the following regression model:
CAR = β0 + β1FERRq + β2Compliance + β3FERRq*Compliance+ β4POST+ β5FERRq*POST + β6POST*Compliance + β7FERRq*Compliance*POST + ΒX + βIndustry FE+ βQuarter-Year FE + βFERRq*X+ βFERRq*Industry FE+
βFERRq*Quarter-Year FE + ɛit (5)
where POST is an indicator variable for firms’ quarterly observations that follow the initial disclosure of compliance with COSO 2013. This variable is then interacted with FERR and
COMPLIANCE to capture whether investors perceive quarterly earnings surprises to be incrementally
more informative for compliance firms relative to noncompliance firms following compliance. The variable of interest in equation (5) is β7, which we predict to be positive.
IV. RESULTS
Descriptive Statistics and Correlations
Table 1 provides descriptive statistics for the variables used in the study. Panel A shows that approximately 91 percent of companies in the sample comply with the COSO 2013
framework in the first fiscal year following implementation. Panel B shows that the mean (median) values of the accounting conservatism measures, C_SCORE, CON_ACC, and CON_SKEWNESS, are 0.402 (0.369), 0.022 (0.012), and 0.457 (0.018), respectively. These
19
values are relatively higher than those found in Ahmed and Duellman (2013) probably due to differences in sample composition and the period under examination. Panel C shows the mean (median) values of the variables used in the short-window earnings response coefficients (ERC).
[Insert Table 1 here]
Table 2 provides the descriptive statistics for the compliance and noncompliance firms separately, and differences in mean and median values between the compliance and
noncompliance firms. We examine mean differences using t-tests and median differences using Pearson chi-square tests. The mean differences in Size, Leverage, KPMG, ICMW, Segments, and M&A are statistically significant under the t-test. These initial results suggest that compliance
firms are larger, more highly leveraged, less likely to be audited by KPMG, and have less material weaknesses in internal control compared to noncompliance firms. Compliance firms also are less complex in that they have fewer operating and geographic segments. However, inconsistent with our expectation, we find that compliance firms are more likely to engage in merger and acquisition activity, and some evidence that compliance firms have higher misstatement risk.
[Insert Table 2 here]
Table 3 presents the Pearson correlation coefficients among the variables. In Panel A, Compliance has a positive and significant (p<0.01) correlation with Size. In contrast, Compliance
is negatively and significantly correlated with KPMG, ICMW, and Segments (p<0.01). Panel B shows that the correlation coefficient between C_SCORE and CON_ACC is positive and
significant (p<0.01). The primary correlation of interest is between Compliance, and C_SCORE and CON_ACC, respectively. As expected, we find a significant positive correlation between two proxies for accounting conservatism, C_SCORE and CON_ACC, and Compliance. Consistent
20
with our expectation, C_SCORE has negative and significant (p<0.01) correlations with Market-to-Book, Sales_Growth, Rd_Adv, and Litigation. We also find positive and significant (p<0.01)
correlation coefficients between C_SCORE, and Size, Leverage, Firm_Age, and Big4,
respectively. On the other hand, CON_ACC is positively and significantly (p<0.01) correlated with Leverage and Rd_Adv, respectively, and negatively and significantly (p<0.01) correlated with Size, ROA, and Firm_Age.
[Insert Table 3 here]
Regression Results
Determinants of Compliance with COSO 2013 Framework
We next discuss our multiple variable regression results. Table 4 shows the results of estimating equation (1) to test Hypotheses 1a through 1h. Consistent with H1a, we find a negative and significant (p<0.001) coefficient on ICMW in columns (2) and (3), suggesting that firms with weaker internal controls are less likely to comply with the COSO 2013 framework. Consistent with H1b, we find that the coefficient on Size is positive and significant (p<0.001) across all three columns, which indicates that larger firms are more likely to comply with the COSO 2013 framework after the transition period. Consistent with H1e, we find that the
coefficient on Segments is negative and significant (p<0.001) in all columns, suggesting that firm complexity delays compliance. In column (3), we find that firms with higher misstatement risk (F_SCORE) are less likely to comply promptly, consistent with H1h. We also find that the coefficient on KPMG is negative and significant (p<0.001) in all columns, consistent with our discussions with a COSO board member suggesting that KPMG was less likely to encourage clients’ prompt compliance with the COSO 2013 framework. We fail to find evidence in support of H1c, H1d, H1f, and H1g.
21
[Insert Table 4 here]
COSO 2013 Framework and Accounting Conservatism
In this section, we investigate the relation between compliance with the COSO 2013 framework and accounting conservatism using our three proxies for accounting conservatism. Table 5 presents the results of the pooled OLS regression examining the conditional
conservatism measure, C_SCORE as the dependent variable. We present the results without inclusion of the inverse Mills ratio in Column (1), and we present the results with the inverse Mills ratio in Column (2). In both columns, we find that the coefficient on Compliance is positive and significant (p<0.001), indicating that firms that comply with the COSO 2013 framework use more conservative accounting relative to noncompliance firms. The sign of coefficients on control variables are consistent across all columns. Although inconsistent with prior research findings (LaFond and Watts 2008; Goh and Li 2011; Ahmed and Duellman 2013), we find a positive and significant (p<0.001) coefficient on Size suggesting that larger firms use more conditionally conservative accounting. Consistent with Ahmed and Duellman (2013), the coefficient on Leverage is positive and significant (p<0.001), indicating that highly leveraged firms have more conservative accounting. We also find a negative and significant (p<0.001) coefficient on Market-to-Book consistent with Ahmed and Duellman (2013). The coefficient on Firm_Age is positive and significant (p<0.001), indicating that older firms exhibit greater
accounting conservatism. We find insignificant associations between C_SCORE and ROA, Sales_Growth, Rd_Adv, Litigation, Big4, and the inverse Mills ratio, suggesting that more
conservative accounting is not associated with the unobservable factors associated with prompt compliance.
22
Table 6 presents the results of the pooled OLS regressions with our second proxy of unconditional conservatism measure, CON_ACC as a dependent variable. Column (1) presents the results without the inverse Mills ratio, and Column (2) presents the results with the inclusion of the inverse Mills ratio. Consistent with Table 5, we find that the coefficient on Compliance is positive and significant (p<0.01) Consistent with the findings of Ahmed and Duellman (2013), we find a negative and significant (p<0.001) coefficient on Size, and a positive and significant (p<0.01) coefficient on Leverage. Unlike findings in Ahmed and Duellman (2013), we find a positive and significant (p<0.001) coefficient on Market-to-Book. We find a negative and
significant (p<0.001) on ROA and on Firm_Age, indicating that that firms with a higher return on assets and that are older use less conservative accounting. The coefficient on Rd_Adv is positive and significant (p<0.1), consistent with Ahmed and Duellman (2013). We also find that the coefficient on Big4 is positive and significant, indicating that firms audited by Big 4 auditors exhibit greater accounting conservatism. In Column (2), we find a negative and significant coefficient on the inverse Mills ratio.
[Insert Table 6 here]
Table 7 presents the results of the pooled OLS regressions examining the unconditional conservatism measure, CON_SKEWNESS as the dependent variable. Consistent with Tables 5 and 6, we find a positive and significant coefficient on Compliance in both columns. We also find that the coefficients on Firm_Age, Rd_Adv, Litigation are negative and significant. The coefficient on Big4 is positive and significant in Column (2) and the coefficient on
Inverse_Mills_Ratio is significantly negative, indicating the importance to control for the
unobservable factors associated with COSO 2013 compliance. [Insert Table 7 here]
23
Table 8 presents the results of the pooled OLS regression using the Basu (1997)
specification for asymmetric timeliness as a proxy for conditional accounting conservatism. We find that the coefficient on D*Return*Compliance is positive and significant (p<0.10). This positive relation suggests that compliance firms tend to engage in more conservative accounting, compared to noncompliance firms. The coefficient on D*Return*Size is negative and significant (p<0.001), suggesting that larger firms use less conservative accounting. We find a positive and significant (p<0.001) coefficient on the coefficient on D*Return*Leverage. This finding
indicates that higher leveraged firms have more conservative accounting. [Insert Table 8 here]
Taken together, the results in Tables 5 through 8 suggest that firms that promptly comply with the COSO 2013 framework provide more conservative financial reporting relative to noncompliance firms. To the extent that conservative accounting reflects strong governance practices and a strong internal control environment, prompt compliance with the updated COSO framework is a means to alleviate agency costs by signaling this commitment to external parties.
Compliance with COSO’s 2013 Framework and Investor Perceptions of Earnings Quality
Table 9 presents the results of the tests of our third hypothesis examining investor perceptions of compliance with the COSO 2013 framework. Specifically, we examine the short-window market response to quarterly earnings surprises. Consistent with prior research, we find a positive and significant earnings response coefficient (FERR) in both columns. With regard to H3, we find that the coefficient on FERR*Compliance in column (1) is positive and significant (p=0.002). We also find a positive and significant coefficient on FERR*Compliance*POST in column (2) (p=0.028). Taken together, these findings suggest that investors find quarterly earnings surprises more informative for firms complying with the COSO 2013 framework, relative
24
to noncompliance firms, and that this informativeness is incrementally higher following initial disclosure of compliance.
V. ADDITIONAL ANALYSES
Subsample Tests with Non-ICMW Firms
Our findings suggest that prompt compliance with the updated COSO framework can help to alleviate agency costs by signaling a commitment to strong internal controls. However, this signaling would be most important for firms without reported material weaknesses. Given that firms with reported material weaknesses in internal control are less likely to promptly comply with the COSO 2013 framework, we examine whether our results are robust to excluding these firms reporting material weaknesses in internal controls from the analyses. Table 10 presents the results of our tests examining the association between compliance with COSO 2013 and
accounting conservatism as well as investor perceptions of the quarterly earnings of compliance firms. Panels A through C present the results using three proxies for accounting conservatism (C_SCORE, CON_ACC, and CON_SKEWNESS). In all three panels, we continue to find a positive and significant coefficient on Compliance. In Panel D, we re-examine our short-window ERC tests after excluding firms reporting material weaknesses in internal controls. We find consistent evidence that even among firms not reporting material weaknesses in internal controls, that investors perceive prompt compliance with the COSO 2013 framework as a signal of a commitment to a strong internal control environment.
[Insert Table 10 here]
Possibility of Omitted Variable Bias
To ensure the robustness of our results, we consider the possibility of omitted variable bias. Following Ahmed and Duellman (2013), we add cash flows from operations (CFO) and
25
volatility of sales (Sales_Vol) and re-estimate the previous equation (2) using three conservatism measures. CFO is measured as cash flows from operations divided by average total assets. Sales_Vol is calculated as the standard deviation of the natural sales between year t-1 and t-5. In
this analysis, we exclude return on assets (ROA) because CFO and ROA are highly correlated and thus it may lead to multicollinearity problems.5 The pooled OLS regression model is as follows:
ACC_Conservatismi,t = β0 + β1Compliancei,t + β2Sizei,t + β3Leveragei,t
+ β4Market-to-Booki,t + β5Firm_Agei,t + β6CFOi,t+ β7Sales_Voli,t + β8Sales_Growthi,t + β9Rd_Advi,t + β10Litigationi,t + β11Big4i,t
+ β12Inverse_Mills_Ratioi,t + Σi,tIndustry_Dummy + ɛi,t, (6)
where ACC_Conservatismrepresents three proxies for accounting conservatism, C_SCORE, CON_ACC, and CON_SKEWNESS and all other variables as previously defined. We continue to
find consistent results with those previously tabulated. Specifically, the coefficients on
Compliance in all columns are positive and significant in Panel A. In Panel B, we find consistent
results when limiting the sample to firms not reporting material weaknesses in internal control. Thus, our primary findings are robust to this alternative model specification.
[Insert Table 11 here]
VI. CONCLUSION
In this study, we examine whether prompt compliance with the COSO 2013 framework provides an indication of a commitment to a strong internal control environment. We first investigate determinants of prompt compliance with the new framework and then perform several analyses to investigate whether prompt compliance is associated with more conservative financial
5 Our result shows that the Pearson correlation between CFO and ROA is about 0.80, indicating that they are highly correlated.
26
reporting. Finally, we examine whether investors perceive compliance with the updated framework as an indication of a firm’s commitment to a strong system of internal control.
We find that compliance is more likely among firms that do not disclose internal control weaknesses, are larger, more highly leveraged, older, less complex, are involved in more merger and acquisition activity, have restructuring charges, and operate in a more litigious industry. We find robust evidence that firms that comply with the COSO 2013 framework provide more conservative financial reporting. Finally, we find that investors find quarterly earnings surprises more informative for firms complying with the COSO 2013 framework, relative to noncompliance firms, and that this informativeness is incrementally higher in the quarters following the initial compliance. These results suggest that prompt compliance provides a signal to market participants about the firm’s control consciousness and governance.
This study contributes to the literature by examining whether prompt compliance with the COSO 2013 internal control framework can serve as a signal of a strong system of internal controls (e.g., Doyle et al. 2007a; Goh and Li 2011) to help alleviate agency costs and highlight the importance of prompt
compliance with future framework updates, new rules, or new regulation involving internal control over financial reporting.
27
REFERENCES
Ahmed, A. S., and B. Billings, R. Morton, and M. Stanford-Harris. 2002. The role of accounting conservatism in mitigating bondholder-shareholder conflicts over dividend policy and in reducing debt costs. The Accounting Review 77: 867–890.
Ahmed, A. S., and S. Duellman. 2007. Accounting conservatism and board of director characteristics: An empirical analysis. Journal of Accounting and Economics 43: 411–437.
Ahmed, A. S., and S. Duellman. 2013. Managerial overconfidence and accounting conservatism. Journal of Accounting Research 51: 1–30.
Ashbaugh-Skaife, H., D. Collins, W. Kinney, and R. LaFond. 2008. The effect of SOX internal control deficiencies and their remediation on accruals quality. The Accounting Review 83: 217–250.
Ashbaugh-Skaife, H., D. Collins, W. Kinney, and R. LaFond. 2009. The effect of SOX internal control deficiencies on firm risk and cost of equity. Journal of Accounting Research 47: 1– 43.Ball, R., and L. Shivakumar. 2005. Earnings quality in UK private firms: Comparative loss recognition timeliness. Journal of Accounting and Economics 39: 83–128.
Ball, R., and L. Shivakumar. 2006. The role of accruals in asymmetrically timely gain and loss recognition. Journal of Accounting Research 44: 207–242.
Basu, S. 1997. The conservatism principle and the asymmetric timeliness of earnings. Journal of Accounting and Economics 24: 3–37.
Beneish, M. D., M. B. Billings, and L. D. Hodder. 2008. Internal control weaknesses and information uncertainty. The Accounting Review 83: 665–703.
Chan, L.H., Chen, K.C., Chen, T.Y. and Yu, Y., 2012. The effects of firm-initiated clawback provisions on earnings quality and auditor behavior. Journal of Accounting and
Economics, 54(2), pp.180-196.
Committee of Sponsoring Organizations of the Treadway Commission (COSO). May 2013. Internal Control–Integrated Framework: Executive Summary.
Dechow, P. M., W. Ge, C. R. Larson, and R. G. Sloan. 2011. Predicting material accounting misstatements. Contemporary Accounting Research 28: 17–82.
Doyle, J., W. Ge, and S. McVay. 2007a. Accruals quality and internal control over financial reporting. The Accounting Review 82: 1141–1170.
Doyle, J., W. Ge, and S. McVay. 2007b. Determinants of weaknesses in internal control over financial reporting, Journal of Accounting and Economics 44: 191–223.
Feng, M., C. Li, S. E. McVay, and H. Skaife. 2015. Does ineffective internal control over financial reporting affect a firm’s operations? Evidence from firms’ inventory management. The Accounting Review 90: 529–557.
Francis, J., and B. Ke. 2006. Disclosure of fees paid to auditors and the market valuation of earnings surprises. Review of Accounting Studies 11 (4): 495–523
Francis, J., D. Philbrick, and K. Schipper. 1994. Shareholder litigation and corporate disclosures. Journal of Accounting Research 32: 137–164.
Garcia, L., M. Juan, B. Garcia Osma, and F. Penalva. 2009. Accounting conservatism and corporate governance. Review of Accounting Studies 14 (1): 161-201.
Ge, W., and S. McVay. 2005. The disclosure of material weaknesses in internal control after the Sarbanes-Oxley Act. Accounting Horizons 19: 137-158.
Ghosh, A., Gu, Z. and Jain, P.C., 2005. Sustained earnings and revenue growth, earnings quality, and earnings response coefficients. Review of Accounting Studies, 10(1), pp.33-57.
28
Ghosh, A., S. Kallapur, and D. Moon. 2009. Audit and non-audit fees and capital market
perceptions of auditor independence. Journal of Accounting and Public Policy 28 (5): 369– 85.
Givoly, D., and C. Hayn. 2000. The changing time-series properties of earnings, cash flows and accruals: Has financial reporting become more conservative? Journal of Accounting and Economics 29: 287– 320.
Givoly, D., C. K. Hayn, and A. Natarajan. 2007. Measuring reporting conservatism. The Accounting Review 82: 65–106.
Goh, B. W., and D. Li. 2011. Internal controls and conditional conservatism. The Accounting Review 83: 975–1005.
Hammersley, J. S., L. A. Myers, and C. Shakespeare. 2008. Market reactions to the disclosure of internal control weaknesses and to the characteristics of those weaknesses under section 302 of the Sarbanes Oxley Act of 2002. Review of Accounting Studies 13: 141-165.
Holthausen, R. W., Watts, R. L. 2001. The relevance of the value-relevance literature for financial accounting standard setting. Journal of Accounting and Economics 31: 3–75.
Jensen, M. C. 1993. The modern industrial revolution, exit and the failure of internal control systems. Journal of Finance 48: 831–880.
Jensen, M. C., and W. H. Meckling. 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics 3: .305-360.
Khan, M., and R. L. Watts. 2009. Estimation and empirical properties of a firm-year measure of accounting conservatism. Journal of Accounting and Economics 48: 132–150.
LaFond, R., and S. Roychowdhury. 2008. Managerial ownership and accounting conservatism. Journal of Accounting Research 46: 101–135.
LaFond, R., and R. L. Watts. 2008. The information role of conservatism. The Accounting Review 83: 447– 478.
Rice, S. C., and D. P. Weber. 2012. How effective is internal control reporting under SOX 404? Determinants of the (non-)disclosure of existing material weaknesses. Journal of
Accounting Research 50: 811-843.
Roychowdhury, S., and R. L. Watts. 2007. Asymmetric timeliness of earnings, market-to-book and conservatism in financial reporting. Journal of Accounting and Economics 44: 2–31.
Watts, R. L. 2003a. Conservatism in accounting part 1: Explanations and implications. Accounting Horizons 17: 207–221.
Watts, R. L. 2003b. Conservatism in accounting, Part II: Evidence and research opportunities. Accounting Horizons 17: 287–301.
Watts, R. L., and J. L. Zimmerman. 1978. Towards a positive theory of the determination of accounting standards. The Accounting Review 53: 112-134.
Zhang, J. 2008. The contracting benefits of accounting conservatism to lenders and borrowers. Journal of Accounting and Economics 45: 27–54.
29
Appendix A Variable Definitions
Variables Description
Panel A: Determinants of Compliance with the COSO 2013 Framework
Compliance An indicator variable equal to one if a firm uses the COSO 2013 internal framework after December 15, 2014 and zero otherwise;
Size The natural logarithm of total asset (Compustat data item AT);
Leverage The sum of long-term debt (Compustat data item DLTT) and short-term debt (Compustat data item DLC) divided by total assets (AT);
Firm_Age The natural logarithm of the number of years the firm has CRSP database; KPMG An indicator variable equal to one if the firm was audited by KPMG during the
current year zero otherwise (Audit Analytics database);
ICMW An indicator variable that takes the value of one if a firm has internal control weaknesses (Audit Analytics database) and zero otherwise;
Going_Concern An indicator variable that takes the value of one if a firm receives going concern opinion from its auditor and zero otherwise (Audit Analytics database) and zero otherwise;
Segments The natural logarithm of the sum of the number of operating and geographic segments (Compustat Segments Database);
Foreign_Operations An indicator variable equal to one if a firm reports foreign operations (Compustat data item FCA) and zero otherwise;
M&A An indicator variable equal to one if a firm is engaged in a merger or acquisition (Compustat data item AQP or AQEPS) and zero otherwise;
Restructuring An indicator variable equal to one if the firm has restructuring charges (Compustat data item RCP or RCEPS) and zero otherwise;
Aggregate_Losses An indicator variable equal to one if earnings before extraordinary items (Compustat data item IB) in years t and t-1 sum to less than zero and zero otherwise;
Litigation Following Francis, Philbrick, and Schipper (1994), we set an indicator variable equal to one if a firm falls in a high litigation risk industry as identified by SIC codes 2833–2836, 3570–3577, 3600–3674, 5200–5961, and 7370;
F_SCORE A fraud risk measure developed by Dechow et al. (2011).
Panel B: Compliance with the COSO 2013 Framework and Accounting Conservatism Proxies for Accounting Conservatism
C_SCORE We use a measure of conditional conservatism, C-Score, developed by Khan and Watts (2009);
CON_ACC Following Givoly and Hayn (2000) and Ahmed et al. (2002), we use a measure of unconditional conservatism, calculated as the net income before extraordinary items (Compustat data item IBC) plus depreciation expense (Compustat data item DP) minus cash flow from operations (Compustat data item OANCF), deflated by average total assets, and averaged over a 3-year period centered on year t, multiplied by negative one;
30
CON_SKEWNESS Following Givoly and Hayn (2000) and Ahmed and Duellman (2013), we use another unconditional conservatism measure, CON_SKEWNESS. This measure is calculated as the difference between cash flow skewness and earnings skewness. The skewness of cash flow (earnings) is defined as (𝑥 − 𝜇)3/𝜎3 where 𝑥 is cash flows (earnings), and 𝜇 and 𝜎 are the mean and standard deviation of cash flows (earnings) over the last five years;
Basu's Specification A Basu’s (1997) asymmetric timeliness measure. Explanatory
Variables
Compliance Same definition as in Panel A;
Size Same definition as in Panel A;
Leverage Same definition as in Panel A;
Market-to-Book The ratio of market value of total assets to book value of total assets;
ROA The operating income before depreciation (Compustat data item OIBDP) divided by total assets (Compustat data item AT);
Firm_Age Same definition as in Panel A; Sales_Growth Same definition as in Panel A;
Rd_Adv Research and development costs (Compustat data item XRD) plus advertising expense divided by sales;
Litigation Same definition as in Panel A;
Big4 An indicator variable equal to one if the firm was audited by a Big 4 auditor during the current year zero otherwise (Audit Analytics database);
Inverse_Mills_Ratio The inverse Mills ratio from equation (1).
NI The earnings before extraordinary items divided by the market value of equity at the beginning of the fiscal year;
DR An indicator variable equal to one if a firm’s return is negative and zero otherwise;
RET The buy-and-hold return over the fiscal year.
Panel C: Compliance with the COSO 2013 Framework and Investor Perceptions of Earnings Quality
CAR The abnormal (i.e., market-adjusted) returns cumulated over days [-1, +1] relative to the quarterly earnings announcement;
FERR The analyst forecast error, measured as the difference between reported quarterly earnings per share and the most recent median consensus analyst earnings forecast, deflated by prior quarter stock price;
absFERR The absolute value of FERR;
Loss An indicator variable if net income for the quarter is less than zero;
Restructure An indicator variable if special items is five percent or more of total assets; DE the ratio of short and long-term debt to total equity; (QTR4), (LnMV), and QTR4 An indicator variable for the last fiscal quarter in the respective year; LnMV The natural log of the market value of equity;
31
STD_Return The standard deviation of market-adjusted buy-and-hold returns over the previous 60 months;
POST An indicator variable for firms’ quarterly observations that follow the initial disclosure of compliance with COSO 2013;
32
Appendix B
Khan and Watts’ C-Score
Khan and Watts (2009) develop a measure of conditional accounting conservatism, C-Score. They estimate of the timelines of good news (G-Score) and bad news (C-Score). Following Khan and Watts (2009), we estimate G-Score and C-Score as follows:
𝑋𝑖 = 𝛽1+ 𝛽2𝐷𝑖 + 𝛽3𝑅𝐸𝑇𝑖+ 𝛽4𝐷𝑖 ∗ 𝑅𝐸𝑇𝑖 + 𝑒𝑖 (1) 𝐺 − 𝑆𝑐𝑜𝑟𝑒 = 𝛽3 = 𝜇1+ 𝜇2𝑆𝐼𝑍𝐸𝑖 + 𝜇3𝑀𝑇𝐵𝑖 + 𝜇4𝐿𝐸𝑉𝑖 + 𝑒𝑖 (2) 𝐶 − 𝑆𝑐𝑜𝑟𝑒 = 𝛽4 = 𝜆1+ 𝜆2𝑆𝐼𝑍𝐸𝑖 + 𝜆3𝑀𝑇𝐵𝑖 + 𝜆4𝐿𝐸𝑉𝑖 + 𝑒𝑖 (3)
where, the subscript i indicates the firm, X is earnings, RET is returns, D is an indicator variable that equals to one when RET<0 and zero otherwise. SIZE is the logarithm of the market value of equity. MTB is the market-book-to ratio, measured as the ratio of market value of equity to book value of equity. LEV is firm leverage, measured as the total debt divided by market value of equity. Substituting 𝛽3 and 𝛽4 derived from equations (2) and (3) into regression equation (1) yield:
𝑋𝑖 = 𝛽1+ 𝛽2𝐷𝑖 + 𝑅𝐸𝑇𝑖∗ (𝜇1+ 𝜇2𝑆𝐼𝑍𝐸𝑖 + 𝜇3𝑀𝑇𝐵𝑖 + 𝜇4𝐿𝐸𝑉𝑖) +𝐷𝑖 ∗ 𝑅𝐸𝑇𝑖 ∗ (𝜆1+ 𝜆2𝑆𝐼𝑍𝐸𝑖 + 𝜆3𝑀𝑇𝐵𝑖 + 𝜆4𝐿𝐸𝑉𝑖)
+(𝛿1𝑆𝐼𝑍𝐸𝑖+ 𝛿2𝑀𝑇𝐵𝑖 + 𝛿3𝐿𝐸𝑉𝑖 + 𝛿4𝐷𝑖∗ 𝑆𝐼𝑍𝐸𝑖 + 𝛿5𝐷𝑖 ∗ 𝑀𝑇𝐵𝑖+ 𝛿6𝐷𝑖 ∗ 𝐿𝐸𝑉𝑖) + 𝑒𝑖 (4) Using annual cross-sectional regressions, we estimate above equation (4). Next, we obtain
G-Score and G-Score from the estimated coefficients from equation (4). In our analysis, we use C-Score as a proxy for a conditional accounting measure.
33
Appendix C
Dechow et al.’s F-score
Dechow et al. (2011) develop a fraud risk measure to capture a firms’ financial statement manipulation. To do this, they use SEC's Accounting and Auditing Enforcement Releases (AAER) database. The F-score is derived from the following equation:
PV = −7.893 + 0.790*RSST+ 2.518*ΔREC + 1.191*ΔINV
+ 1.979*SOFT_ASSETS+ 0.171*ΔCS− 0.932*ΔROA + 1.029*ISSUE
where:
RSST = (ΔWC + ΔNCO + ΔFIN)/Average Total Assets, where Δ is the change operator, WC = (Current Assets − Cash and Short-Term Investments) – (Current Liabilities − Debt in Current Liabilities); NCO = Total Assets − Current Assets –
Investments and Advances – (Total Liabilities − Current Liabilities − Long-Term Debt); FIN = (Short-Term Investments + Long-Term Investments) − (Long-Term Debt + Debt in Current Liabilities + Preferred Stock);
ΔREC = ΔAccounts Recevables/Average Total Assets; ΔINV = ΔInventory/Average Total Assets;
SOFT_ASSETS = (Total Assets – PP&E – Cash and Cash Equivalent)/Total Assets; ΔCS = percentage change in cash sales, where cash sales = Sales − ΔAccounts Recevables;
ΔROA = change in return on assets, where return on assets = Net Income/Total Assets; ISSUE = an indicator variable that equals to one if the firm issued securities during the current period and zero otherwise.
Using the above PV, we calculate the F_SCORE as follows: (ePV/ (1+ePV))/ (0.0037), where e indicates exponential function.
